Construction and Building Materials 45 (2013) 289–297

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Magnesium hydroxide, seawater and olive mill wastewater to reduceswelling potential and plasticity of bentonite soil

0950-0618/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.03.053

⇑ Corresponding author. Address: Department of Geodynamics, University ofGranada, Campus de Fuentenueva s/n, 18071, Granada, Spain. Tel.: +34 647101415.

E-mail address: cgunieto@ugr.es (C. Ureña).

C. Ureña a,⇑, J.M. Azañón a,b, F. Corpas c, F. Nieto d, C. León e, L. Pérez c

a Department of Geodynamics, University of Granada, Granada, Spainb Andalusian Institute of Earth Sciences, UGR-CSIC, Granada, Spainc Department of Chemical Engineering and Materials, University of Jaen, Jaen, Spaind Department of Mineralogy and Petrology, University of Granada, Granada, Spaine Isolux Corsan-Corviam Construction, Engineering Office, Madrid, Spain

h i g h l i g h t s

�We prepare samples of clayey soil mixed with olive mill wastewater, mg-hydroxide and seawater.� We study the changes in geotechnical and mineralogical properties of the soil.� The high swelling potential of clayey soil diminishes after treatment.� Some of the treatments could make the soil be suitable for use in construction.

a r t i c l e i n f o

Article history:Received 7 February 2012Received in revised form 6 March 2013Accepted 15 March 2013

Keywords:Magnesium hydroxideSeawaterOlive mill wastewaterBentoniteChemical stabilisationExpansive soil

a b s t r a c t

The improvement in physical properties of expansive soils after addition of non-conventional additives isstudied. The tested agents were magnesium hydroxide, seawater and olive mill wastewater. The use ofthese materials, which can be obtained from natural sources or as by-products derived from industrialprocesses, can lead to greater sustainability of the construction process. In the case of olive mill waste-water, its management is still an issue in many regions where production of olive oil is a main economicactivity. The untreated soil used was a sample of pure bentonite. To evaluate the effects of the additives,the physical properties of the soil such as compaction, consistency, bearing capacity and swelling pres-sure were studied. The mineral compositions of the treated soils were evaluated by XRD tests. Test resultsshowed that the non-conventional additives tested reduced the plasticity and the swelling potential ofthe soil. Indeed the tested agents proved to be very effective to produce reductions in the swelling pres-sure of 60–87% in comparison with the original swelling pressure of the untreated bentonite. X-ray dif-fraction tests proved that magnesium hydroxide, seawater and olive mill wastewater promoted mineralchanges within bentonite, especially smectite to illite conversion. This conversion is due to the alterationspromoted by the additives on conditions such as pH and concentration of different cations (magnesium,calcium, potassium, etc.) The mineral modifications occurred are behind the sharp reductions in swellingpressure and plasticity of bentonite.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The use of natural resources and industrial by-products in civilworks can lead to a more sustainable construction process. Espe-cially, the use of non-conventional additives for soil stabilisationcan offer an interesting alternative to reduce the damaging effectsof expansive clays. In recent years, many authors have tested theeffectiveness of a wide range of alternative stabilisation agents,from biomass to recycled materials [1–11]. Some studies focus

on the mechanical behaviour of the treated soils whilst some otherfocus on the evolution of the microstructure, studying the electri-cal conductivity, mineralogy or formation of cementitious gels.However, and probably due to the difficulty of quantification ofthe mineral composition of fine-grained soils, not much is yetestablished about the relationship between changes in the mineralcomposition of the soil and the evolution of its physical properties,especially when non-calcium based additives are used.

The clarification of this aspect is crucial when testing new addi-tives, since the swelling potential and plasticity of the soil dependon the minerals present in the soil, the montmorillonite being themost expansive [12]. According to Drief et al. [13], the smectite toillite conversion is controlled by temperature, pH and concentra-

Table 1Properties of bentonite.

Chemicalcomposition (%)

Physical properties

SiO2 50.23 Standard proctor maximum dry density (Mg/m3) 0.91Al2O3 7.21 Standard proctor optimum moisture content (%) 66.10Fe2O3 2.17 California Bearing Ratio (CBR) 1.3K2O 1.54 Liquid limit (%) 294.9MgO 17.77 Plasticity index (%) 250.9Na2O 2.36 Swelling pressure (kPa) 220TiO2 0.42 pH 9.5MnO 0.04 Particule size (lm) 75P2O5 0.06

290 C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297

tion of potassium. According to Elert et al. [14], the hydroxidessuch as KOH and NaOH are more efficient than Ca(OH)2 when itcomes to destroying clay minerals.

Petry and Little [15] presented a thorough review on the use ofcalcium-based additives for soil stabilisation: lime, cement and coalfly ash. Nowadays It is known that the most significant of these tra-ditional binders, lime, promotes a quick effect called flocculation-agglomeration of particles and a slow pozzolanic reaction whichis responsible for the strength of lime-treated soils [16]. However,some authors found lime to pose serious disadvantages for soil sta-bilisation in regions with great seasonal variations, marly soils, etc.[17,18,11]. Hence the importance of testing new additives.

But which specific properties are required for a potential stabi-lisation agent? Based on the information provided by Solanki andZaman [16], we would search for materials with a high amountof free lime and/or pozzolanic potential, i.e., alkalinity and highpercentages of silica and alumina [19]. Such an additive would im-prove the mechanical behaviour of a cohesive soil. But according toWhite [12] and Drief et al. [13], the swelling potential and plastic-ity of the soil depends strongly on the presence of expansive clayminerals within the soil.

Xeidakis [20,21] studied the effects of a solution of magnesiumhydroxide on a suspension of swelling clays, finding that the swell-ing potential of clays was reduced. The effects of the magnesiumhydroxide in a powder form when it is added to a soil followingconventional mixing techniques deserves to be further studied.The same can be said on the effects of seawater and olive mill

Fig. 1. XRD pattern

wastewater. In the existing literature there are studies reportingthe effects of the sodium chloride [22–28] and olive cake residue[29,30] on the physical properties of expansive soils. The chemicalcompositions of both seawater and olive mill wastewater involveelements such as calcium, potassium and magnesium which couldpromote a cation exchange when added to clay minerals. Hencethe interest of a further research on the effect of this agents onthe expansive soils.

This paper investigates the effects of non-conventional addi-tives such as magnesium hydroxide, seawater and olive mill waste-water on a sample of bentonite, a clay of high plasticity and highswelling potential. The study is focused on the mineral changespromoted by the additives and their influence on the swelling po-tential and plasticity of bentonite. The main physical properties ofsoil, such as compaction, consistency, bearing capacity and swell-ing pressure are investigated. X-ray diffraction tests were carriedout to analyse the evolution of the mineral composition ofbentonite.

2. Materials

2.1. Bentonite

The soil used in this study was a sample of bentonite. This soil was chosen forbeing a well known pure clay. It is a highly expansive material which undergoes sig-nificant volume changes when the environmental moisture varies.

The Bentonite was analysed in the laboratory, and found to have poor engineer-ing properties. It was characterised for its low-load bearing capacity, high swellingpotential and plasticity.

X-ray diffraction (XRD) tests were conducted to analyse the mineral composi-tion of the expansive soil. The main mineral constituent of the bentonite used inthis study was Montmorillonite with small amounts of other clay minerals andquartz. Montmorillonite is a clay mineral within the group of smectites. The mostexpansive clay minerals belong to this group.

Table 1 presents the engineering properties and chemical composition of theexpansive soil, while Fig. 1 shows the X-ray diffraction pattern of the untreated soil.Two types of montmorillonite were found to be present in the bentonite, with theirfirst peak located at 2-theta angles of 5.895� and 7.076� respectively.

2.2. Magnesium hydroxide

Magnesium hydroxide (Mg(OH)2), which can be found in nature in the form ofbrucite, is also obtained nowadays from certain industrial by-products[10,11,20,21,31]. The magnesium hydroxide used in this study was a very purecommercial type supplied in a powder form. Previous works showed the reductionof swelling potential of clays, when introducing magnesium hydroxide solutions in

of bentonite.

Table 2Chemical composition of seawater according to Hodge [33].

Element Concentration (mol/kg) Element Concentration (mol/kg)

Cl 0.546 N 3.0 � 10�5

Na 0.468 Li 2.5 � 10�5

Mg 0.0532 Rb 1.4 � 10�6

S 0.0282 Mo 1.1 � 10�7

Ca 0.0103 Ba 1.0 � 10�7

K 0.0102 V 3.0 � 10�8

Br 8.4 � 10�4 As 2.3 � 10�8

Sr 9.0 � 10�5 Al 2.0 � 10�8

F 6.8 � 10�5 U 1.4 � 10�8

� Compounds with concentration >1 � 10�8.

Table 3Chemical composition of OMW according to Garcia Tortosa et al.[36].

pH = 5.38

P (g kg�1) 0.8 Cu (g kg�1) 0.022K (g kg�1) 10.4 Mn (g kg�1) 0.056Ca (g kg�1) 8.0 Zn (g kg�1) 0.017Mg (g kg�1) 3.1 Pb (g kg�1) 0.004Na (g kg�1) 0.3 Cr (g kg�1) 0.019S (g kg�1) 1.1 Ni (g kg�1) 0.055Fe (g kg�1) 2.4 Cd (g kg�1) nd

C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297 291

clay suspensions [20,21]. In this study, a powder of magnesium hydroxide wasadded to the soil following conventional mixing techniques to evaluate the effectsof this agent on the expansive soil in the presence of water and emulate the realconditions of stabilisation process in construction sites.

2.3. Seawater

The widespread availability of seawater makes it very interesting for its bene-ficial use in construction activities, always searching for a more sustainable con-struction process. To carry out this study, natural seawater from theMediterranean sea was collected in Malaga, Southern Spain coast.

Bruland [32] carried out a thorough review on the chemical composition of sea-water. More recently, Hodge [33] published the average values of the elements con-tained in this natural resource. According to these previous works, the mainelements which can be found within seawater are sodium (Na+) and chlorine(Cl�), but the presence of other ions such as calcium, magnesium and potassiumis not negligible. Table 2 shows the main values of the chemical composition of sea-water according to Hodge [33].

2.4. Olive mill wastewater

The olive mill wastewater (OMW) is a by-product of the olive oil industry. It isgenerated in centrifuges during the production of olive oil in the so-called two-phase extraction system. OMW used in this study, which had a humidity of 62%,was produced in Jaen (SE Spain) where the olive grove covers a total extension of5900 km2. Previous studies on the characterisation of OMW concluded that it hasa high content of potassium [34–36]. Table 3 shows the results of the study carriedout by Tortosa et al. [36] which established the chemical composition of the solidphase of the same OMW used in this study.

Fig. 2. Pictures of lab equipment: (a) preparation of swelling pres

3. Methodology

3.1. Preparation of samples

In this work, the untreated bentonite soil was firstly character-ised. Afterwards, several non-conventional stabilisation agentswere added to the original untreated bentonite soil: magnesiumhydroxide, seawater and olive mill wastewater (OMW). The mix-tures were prepared with three different dosages of additive (5–10–15%).

All mixes were prepared under the same conditions in a labora-tory with a controlled temperature of 21 ± 1� C. After weighing theexact quantities of dry soil and dry additive, they were placed in amixing tray and thoroughly mixed for at least 10 min. Water wasthen added to the mix in order to reach the optimum moisture con-tent calculated for the original untreated soil (66%). The soil, addi-tive and water were mixed in an industrial mixer for at least tenmore minutes. Afterwards, the mixes were placed in a curing roomfor the required length of time. The curing room had the followingconditions: temperature of 21 ± 1�C, humidity of 95 ± 3%.

Two experiments were conducted. On one hand, all the speci-mens made were tested to determine their engineering properties.On the other hand, X-ray diffraction (XRD) tests were developed,obtaining the XRD pattern of the samples. All the tests were carriedout at two curing times: 15 days and 30 days.

3.2. Geotechnical tests

To determine the engineering properties of the original soil andthe soils prepared in the laboratory, tests were carried out to ob-tain compaction properties, consistency limits, bearing capacityand swelling potential. Fig. 2 shows pictures of the equipmentused: (a) soil and cell for swelling pressure test; (b) Proctor com-paction moulds; (c) oedometers.

3.2.1. Standard Proctor testThe samples of treated soil were subjected to the standard Proc-

tor test at 15 and 30 days of curing to compare the results withthose obtained for the original untreated soil.

The standard Proctor tests were carried out according to theSpanish standard UNE 103500 [37]. Following this standard, sev-eral samples of soil are to be prepared at different moisture con-tents. The sample is compacted in a mould using a hammer of2.5 kgs. The test must be repeated for samples with different mois-ture contents which promotes a variation of the compaction degreeachieved in the mould. A curve can be plotted which shows therelationship between moisture of the samples (%) and density ofthe samples (Mg/m3). The maximum value of density is calledmaximum dry density (MDD). The moisture at which MDD isachieved is the optimum moisture content (OMC).

sure cell; (b) molds for standard Proctor test; (c) oedometers.

Fig. 3. Swelling pressure test apparatus, according to Akcanca and Aytekin (2012).

Fig. 4. Maximum dry density of treated samples after 30 days of curing.

Fig. 5. Optimum moisture content of treated samples after 30 days of curing.

292 C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297

3.2.2. Atterberg limitsThe Atterberg consistency limits were determined in accor-

dance with the Spanish standards. The standards UNE 103103[38] and UNE 103104 [39] present a methodology to obtain the li-quid limit (LL) and plastic limit (PL) respectively. Plasticity index(PI) is obtained according to the following formula: PI = LL–PL.

3.2.3. California bearing ratioCalifornia Bearing Ratio (CBR) test on stabilised soil specimens

was conducted as per Spanish standard UNE 103502 [40]. To carryout the CBR test, the specimens were first assessed under a Modi-fied Proctor Test in accordance to the standard UNE 103501 [41].The specimens were then cast into the CBR-mold with the samecompactive energy per volume as in the Modified Proctor Test.

3.2.4. Swelling pressure testSpanish Standard UNE 103602 [42] was followed to determine

the swelling pressure of the original untreated soil and all the mix-tures prepared in this study. Both the samples of untreated andtreated soil for the swelling pressure tests were prepared at theoptimum moisture content calculated by the Standard Proctor test(OMCbentonite = 66.1%).’’

To carry out the swelling pressure test, the same equipmentused for consolidation tests is required. The objective of this testis measuring the pressure exerted by the soil on the upper layersdue to a swelling phenomenon. For that, a sample of soil is placedin the consolidation cell between two porous stones and confinedlaterally using several collars. Afterwards, the sample is subjectedto saturation. The vertical strain of the specimen must be pre-vented during the whole test. If the initial saturation promotesan upward movement, the seating pressure applied must be in-creased to hold the reading of the gauge at the initial value. Finally,the swelling pressure value (Ph) is obtained by the following for-mula: Ph = (Q/S) � 103 (kPa), where Q is the equilibrium loadingand S is the section area of the sample. For low permeability soils,the equilibrium loading can be reached after at least 24 h oftesting.

A picture of the cell and oedometer is depicted in Fig. 2. Fig. 3shows a diagram of the swelling pressure test apparatus accordingto Akcanca and Aytekin [43].

3.3. Mineralogy: X-ray diffraction

In this study, X-ray diffraction (XRD) techniques were used toassess the evolution of clay minerals after the treatments focusingon the intensity and position of the first peak of montmorillonite.

XRD patterns were obtained using the powder method (totalfraction) and preparing oriented aggregates (fine fraction). The ori-ented aggregates are usually divided in two subsamples, one ofwhich can be subjected to ethylene–glycol (EG) vapor exposurewhich allows to characterise more accurately the first peak ofmontmorillonite. All the samples prepared were analysed with aPhilips X’Pert–MPD diffractometer using an anticathode Cu Ka at45 kV and 50 mA. A scan rate of 2�(2h)/min was used for the pow-

der samples over the range 2–70�(2h) while the scan rate for ori-ented aggregates was 1�(2h)/min. In this study, the identificationof the mineral phases in the soils prepared was made by usingthe specific software Xpowder 2004 [44].

4. Results and discussion

4.1. Effect on compaction

The values of maximum dry density (MDD) and optimum mois-ture content (OMC) in the treated samples are plotted in Figs. 4 and5 respectively.

In this study, a relationship was found to exist between theMDD of the sample and the percentage of additive present (linearcorrelation factor R2 = 0.88). The values of MDD obtained at oldestsamples with the highest dosage of additive were 1.09 Mg/m3 (15%Mg(OH)2, 1.09 Mg/m3 (15% seawater) and 1.03 Mg/m3 (15% olivemill wastewater). Those values represent an increase over the ori-ginal value (0.91 Mg/m3). This increase in the MDD can be attrib-uted to changes in the grain size since the additives used in thisstudy had not high unit weights but they could promote changesin the particle size distribution.

Fig. 5 shows that addition of magnesium hydroxide, seawater orolive mill wastewater yields to decreases in the OMC. There are

Table 4Results of Atterberg Limits tests.

Combination 1st Stage of curing 2nd Stage of curing

5% 10% 15% 5% 10% 15%

Liquid limit (%)Mg-hydroxide 216.2 193 169.1 283.5 199.4 177.6Seawater 253.7 234.7 206.6 238.6 215.6 193.9OMW 211.5 199 178.9 205.5 197.6 163.5

Plastic limit (%)Mg-hydroxide 46.9 48.9 42 44 58.9 53.1Seawater 35.6 34.4 33 39 30.9 32.6OMW 41.5 42.7 43.7 41 46.8 34.6

Plasticity index (%)Mg-hydroxide 169.3 144.1 127.1 239.5 140.5 124.5Seawater 218.1 200.3 173.6 199.6 184.7 161.3OMW 170 156.3 135.2 164.5 150.8 128.9

C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297 293

studies in which the addition of fly ash [45] or rice husk ash [46]promoted an increase of the optimum moisture content, probablydue to the porous properties of ashes and the increasing amount ofwater required to achieve MDD when soil is mixed with finegrained materials [47,48].

4.2. Effect on consistency

The results of the Atterberg limits tests carried out for soil trea-ted with magnesium hydroxide, seawater and olive mill wastewa-ter are summarised in Table 4.

The addition of magnesium hydroxide to bentonite clay rapidlyproduced a sharp decrease in PI from 250.9 to values under 170, asdisplayed in Fig. 6. The PI values were proved to maintain a de-crease tendency during the second stage of curing, decreasingslowlier than in the first stage.

The addition of seawater to the sample of bentonite also yieldeda decrease in PI of treated soils, although the final value obtainedwere not as low as the values achieved by adding magnesiumhydroxide or olive mill wastewater (Fig. 6). This reduction of plas-ticity of the soil is consistent with the results of the study carriedout by Horpibulsuk [49] which show a reduction of plasticity in

Fig. 6. Evolution of plasticity index in treated samples:

clays when the percentage of sodium chloride increases. Accordingto Horpibulsuk [49], the reduction of plasticity is due to the com-pression of the diffuse double layer of kaolinite particles.

As seen in Table 4 and Fig. 6, the addition of olive mill wastewa-ter (OMW) to the soil reduced the plasticity of the soil. Regardingthe initial value of PI of bentonite (250.9), dosages of 5%, 10% and15% of OMW produced reductions of 34%, 40% and 48% over the to-tal value of the plasticity index respectively.

Given the different plasticity of soils depending on their mineralcomposition [12], the reduction of plasticity promoted by magne-sium hydroxide, seawater and OMW can be attributed to the alter-ation of the pH and the alteration in the concentration of calcium,magnesium and potassium, etc. which leads to transformationsfrom expansive phases (smectite) to non-expansive clay minerals(such as illite). In addition, as explained by Xeidakis [20], theadsorption of magnesium hydroxide by clays is similar to theadsorption of calcium hydroxide. This could have promoted a coar-ser particle size distribution and hence a reduction in plasticity inthe samples treated with magnesium hydroxide.

4.3. Effect on bearing capacity

In general, the treatments carried out in this work produced anincrease in the CBR values of the samples (Fig. 7). This is becausethe addition of magnesium hydroxide, seawater and olive millwastewater produces a flocculation of particles, due to the varia-tions of pH and the cation exchange induced in clays. As a result,the physical properties of bearing capacity and compaction aremodified. Indeed a relationship was found to exist between theCBR and the MDD of samples in this study, with a linear correlationfactor R2 = 0.72. This means that, to some extent, the changes inCBR were related with the compaction properties of the soil.

In the first stages of curing, low dosages (5%) of seawater andOMW produced a decrease in the CBR of the bentonite, whichcan be partially attributed to the formation of voids in the soil asa consequence of the flocculation of particles, without achievinga well graduated particle size distribution. For the rest of dosages,the CBR increased from the first stages of curing. Within curingtime, the continuation of the flocculation can produce the readjust-

(a) magnesium hydroxide; (b) seawater; (c) OMW.

Fig. 7. Evolution of CBR in treated samples: (a) magnesium hydroxide; (b) seawater; (c) OMW.

294 C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297

ment of particles leading to a more compacted sample and hence asample with higher bearing capacity.

The results obtained with OMW were slightly lower than thoseobtained with magnesium hydroxide and seawater. Given magne-sium hydroxide and seawater provide a more alkaline environ-ment, this could be behind this difference.

In Spanish standard PG-3 [50], a minimum CBR of 3% is requiredto classify the soil as tolerable for its use in core of embankments.The initial California bearing ratio (CBR) of bentonite clay was 1.3%.After 30 days, the CBR values of all the samples prepared werehigher than the initial one. The increase of CBR due to the additionof 10% and 15% of magnesium hydroxide and seawater was signif-icant enough as to change the classification of the soil to tolerable.

Fig. 8. Evolution of swelling pressure in treated samples

4.4. Effect on swelling pressure

Given the undoubted necessity of reducing the swelling poten-tial of expansive soils, these results are the most significant of thisstudy. As a first statement, it can be claimed that magnesiumhydroxide, seawater and olive mill wastewater produced, whenadded to the expansive soil, a dramatic decrease in the swellingpressure of the bentonite. The evolution of swelling pressure ispresented in Fig. 8.

The initial swelling pressure of the original bentonite was220 kPa. After the addition of magnesium hydroxide, the swellingpressure obtained was 90 kPa (for 5% additive), 75 kPa (for 10%additive) and 40 kPa (for 15% additive). This means reductions of59–81% in the swelling pressure of the soil. As it can be seen in

: (a) magnesium hydroxide; (b) seawater; (c) OMW.

Fig. 9. XRD pattern of untreated bentonite and samples treated with 15% dosage of additives. Evolution of first peak of montmorillonite.

C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297 295

Fig. 8, the addition of different percentages of seawater and OMWproduced reductions of the swelling pressure of bentonite rangingfrom 68% to 75% in the case of seawater and from 73% to 89% in thecase of OMW. The final values of swelling pressure achieved for theOMW-treated samples were between 25 and 60 kPa, depending onthe dosage of additive, and between 55 and 70 kPa in the case ofsamples treated with seawater.

Using up to 3% of lime, Akcanca and Aytekin [43] achievedreductions of 78% in swelling pressure of an expansive soil madeof sand and bentonite, leading to final values of swelling pressureas low as 30 kPa. Despite the fact that the dosages of lime used inthe mentioned work were low, these results are comparable to theresults presented in this paper, especially taking into account thatthe expansive soil tested in the mentioned study had 50–80% con-tent of sand which do not contribute to the swelling potential.

The reasons for such reductions in the swelling pressure of thetreated samples are related to the weak stability of montmorillon-ite. Elert et al. [14] studied the destruction of both expandable andnon-expandable clay minerals in alkaline environments. Further-more, Drief et al. [13] pointed out that montmorillonite to illiteconversion is easily enabled. Both magnesium hydroxide and sea-water alter the pH of the soil, whilst OMW increases the concentra-tion of potassium, both of them being main parameters to destroymontmorillonite, whose presence is responsible for the swellingpotential of expansive soils. Therefore, all the additives testedhad accurate properties to produce changes in the mineral compo-sition of bentonite, and those changes are behind the reduction ofswelling pressure of treated samples.

4.5. Effect on mineralogy

The effects of the additives tested in this study on the mineralcomposition of bentonite have been observed by X-ray diffraction(XRD) tests. Fig. 9 shows the pattern of treated samples with 15% ofeach additive in comparison with the XRD pattern of bentonite. Inthe pattern of bentonite, two different types of montmorillonitewith their first peaks located at 2h angle of 5.895� and 7.076� wereidentified as calcium montmorillonite and sodium montmorillon-ite, respectively.

According to Moore and Reynolds [51], the name of sodium, cal-cium or magnesium montmorillonite refers to the main interla-

mellar cation. The cation exchange capacities (CECs) ofmontmorillonite is high, and expansion takes place as water orsome polar inorganic compound enters the interlayer space [51].

It can be seen that in the treated samples, regardless of theadditive used, these two peaks have turned into just one locatedat 2h angle of 7.281�. This peak, identified as the first peak of amontmorillonite, was not only shifted to the right in the patternbut it also had lower intensity of reflection. According to Mooreand Reynolds [51], this entails a reduction in the amount of mont-morillonite present in the sample along with a reduction in the d-spacing from 12.482 Å and 14.980 Å to just 12.15 Å.

The presence of imbalance electrical charge, sodium based clayminerals and CEC constitutes the swelling nature of bentonite [52].Magnesium hydroxide, seawater and OMW changed the pH of themedium. Magnesium hydroxide and seawater provided magne-sium and calcium. OMW altered the concentration of potassium.Therefore, this shift of the first peak of montmorillonite can beattributed to the exchanges taken place among different cations,namely Ca2+, Mg2+, Na+ or K+. This exchange can produce a reduc-tion in the interlamellar spacing leading to more stable structures.

Table 5 shows the exact position of the first peak of montmoril-lonite in all the samples treated with magnesium hydroxide, sea-water and OMW. The same effects previously described (lowerintensity of reflection and right-shift of first peak of montmorillon-ite) can be observed for all the treated samples.

Therefore, on the basis of these results, it can be stated that thetreated samples proved to have a lower amount of montmorillon-ite, and the montmorillonite present proved to have lower expan-sion potential. The additives used produced a noticeable change inthe mineral composition of bentonite, tending to a reduction ofmontmorillonite present in the sample. Since montmorillonite isresponsible for the swelling potential but also for the high plastic-ity of the soil, whilst illite is a non-expansive mineral, thesechanges let us understand the reductions in plasticity and swellingpressure observed in the treated samples and addressed in previ-ous sections.

4.6. Summary of results and comparison of additives

The non-conventional additives tested in this study (magne-sium hydroxide, seawater and OMW) produced similar effects on

Table 5Evolution of first peak of montmorillonite in treated samples.

Sample Intensity (counts) d-Spacing (Å) 2-Theta angle (�)

Untreated samples

BentoniteSodium montmorillonite 802 12.48280 7.076Calcium magnesium montmorillonite 591 14.97996 5.895

Treated samplesBentonite + 5% Mg(OH)2 305 11.96278 7.384Bentonite + 10% Mg(OH)2 372 12.13123 7.281Bentonite + 15% Mg(OH)2 406 12.21725 7.230Bentonite + 5% seawater 461 12.11074 7.293Bentonite + 10% seawater 512 12.10050 7.281Bentonite + 15% seawater 470 12.18410 7.281Bentonite + 5% OMW 508 12.12820 7.281Bentonite + 10% OMW 580 12.13123 7.281Bentonite + 15% OMW 522 12.07290 7.332

Fig. 10. Comparison of additives in terms of plasticity and swelling pressure of thetreated soils.

296 C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297

the bentonite. First, a flocculation of particles responsible for thechanges in the compaction and bearing capacity of the soil. Second,a change in the mineral composition of bentonite, tending to de-crease the amount of montmorillonite present in the sample. Thealteration of the original conditions of formation of montmorillon-ite can easily enable the smectite to illite conversion. As stated byDrief et al. [13], this conversion is mainly controlled by the pH,temperature, curing time and access to potassium.

Magnesium hydroxide and seawater were able to alter the pH ofthe medium and provide cations for the cation exchange. Further-more, OMW contained a high concentration of potassium. All theseconditions altogether are behind the changes occurred in the min-eral composition of bentonite. And the influence of the mineralcomposition on the physical properties of the soil was remarkable,since the reduction of montmorillonite was accompanied by asharp decrease in the plasticity and swelling pressure of the trea-ted samples.

In terms of comparison, the results obtained by addition ofmagnesium hydroxide, seawater and OMW were very similar, asseen in Fig. 10. The final values of plasticity index achieved bytreatment with magnesium hydroxide were slightly better thanthose obtained with the other additives. In terms of swelling pres-sure, OMW led to the greatest reductions. The addition of seawateralso produced significant improvements in the plasticity andswelling potential of the soil. Hence, all the additives tested provedto have promising properties to reduce the plasticity and swelling

potential of the additives while improving other physical proper-ties such as compaction and bearing capacity. The effects of differ-ent combinations of these additives on the expansive soil deserve afurther investigation.

5. Conclusions

In this study, different dosages of magnesium hydroxide, sea-water and olive mill wastewater were added to a sample of ben-tonite. On the basis of the results of the geotechnical and mineraltests, the following conclusions can be drawn:

– The results of this study showed that the high swelling pressureand plasticity index of bentonite depend on its mineral compo-sition, i.e., the amount of expansive phases such as montmoril-lonite present in the sample. Changes in the intensity andposition of the first peak of montmorillonite and illite in thesamples correspond with noticeable changes in plasticity andswelling pressure.

– The addition of magnesium hydroxide, seawater or olive millwastewater to the bentonite promotes a reduction in theamount of montmorillonite present in the soil, partially due tothe cation exchange capability enabled by introduction of cat-ions (magnesium, potassium, calcium, sodium, etc.).

– In addition, the variation of the original conditions in whichmontmorillonite was formed can easily enable the montmoril-lonite to illite conversion. After addition of magnesium hydrox-ide, seawater or olive mill wastewater to the bentonite, this soilsuffers sharp decreases in swelling pressure and plasticityindex, due to the alterations promoted by the additives on con-ditions such as pH and concentration of different cations (mag-nesium, calcium, potassium, etc.)

– Along with the improvement on plasticity and swelling poten-tial, the non-conventional additives tested in this study pro-duced increases of maximum dry density and Californiabearing ratio (CBR) of the bentonite, whilst the optimum mois-ture content of the treated samples was always inferior to thatof the original bentonite. The samples treated with at least 10%of additive reached CBR index above 3%, which is the valuerequired for Spanish standard PG-3 to use a soil in core ofembankments [50].

Acknowledgements

This study was supported by the research projects CGL2011-29920 and TOPOIBERIA CONSOLIDER-INGENIO2010 (SpanishMinistry of Science and Innovation). The authors want to thank

C. Ureña et al. / Construction and Building Materials 45 (2013) 289–297 297

Corsan Corviam S.A. (Isolux Group) and Lidycce S.L. for their fund-ing and support. The authors also want to offer special thanks toMr. Patrick Hanrahan (Imperial College London, UK) and Mr. TobiasNeville (University College London, UK) for their invaluable helpand advice.

References

[1] Clifton JR, Brown PW, Frohnsdorff G. Uses of waste materials and by-productsin construction. Part I Resour Recov Conserv 1980;5:139–60.

[2] Mathur VK. Composite materials from local resources. Constr Build Mater2006;20:470–7.

[3] Attom MF. Shear strength characteristics of Irbid clayey soil mixed with ironfilling and iron filling-cement mixture. Environ Geol 2007;55(4):781–8.

[4] Hossain KMA, Lachemi M, Easa S. Stabilized soils for construction applicationsincorporating natural resources of Papua new Guinea. Resour Conserv Recy2007;51:711–31.

[5] Degirmenci N. The using of waste phosphogypsum and natural gypsum inadobe stabilization. Constr Build Mater 2008;22:1220–4.

[6] Eren S, Filiz M. Comparing the conventional soil stabilization methods to theconsolid system used as an alternative admixture matter in Isparta Daridetematerial. Constr Build Mater 2009;23:2473–80.

[7] Peethamparan S, Olek J, Diamond S. Mechanism of stabilization of Na-montmorillonite clay with cement kiln dust. Cement Concrete Res2009;39:580–9.

[8] Hossain KMA, Mol L. Some engineering properties of stabilized clayey soilsincorporating natural pozzolans and industrial wastes. Constr Build Mater2011;25:3495–501.

[9] Rahmat MN, Ismail N. Sustainable stabilisation of the Lower Oxford Clay bynon-traditional binder. Appl Clay Sci 2011;52:199–208.

[10] Seco A, Ramirez F, Miqueleiz L, Garcia B. Stabilization of expansive soils for usein construction. Appl Clay Sci 2011;51:348–52.

[11] Seco A, Ramirez F, Miqueleiz L, Garcia B, Prieto E. The use of non-conventionaladditives in Marls stabilization. Appl Clay Sci 2011;51:419–23.

[12] White WA. Atterberg plastic limits of clay minerals. Urbana-Champaign: University of Illinois; 1949.

[13] Drief A, Martinez-Ruiz F, Nieto F. Velilla Sanchez N. Transmission electronmicroscopy evidence for experimental illitization of smectite in K-entichedseawater solution at 50 �C and basic pH. Clay Clay Miner 2002;50(6):746–56.

[14] Elert K, Sebastian E, Valverde I, Rodriguez-Navarro C. Alkaline treatment ofclay minerals from the Alhambra Formation: Implications for the conservationof earthen architecture. Appl Clay Sci 2008;39:122–32.

[15] Petry TM, Little DN. Review of stabilization of clays and expansive soils inpavements and lightly loaded structures—history, practice, and future. J MaterCiv Eng 2002;14(6):447–60.

[16] Solanki P, Zaman M. Microstructural and mineralogical characterization of claystabilized using calcium-based stabilizers. In: Kazmiruk V, editor. Scanningelectron microscopy. Rijeka: Intech; 2012. p. 771–98.

[17] Raymond NY, Ouhadi VR. Experimental study on instability of bases on naturaland lime/cement stabilized clayey soils. Appl Clay Sci 2006;35(3–4):238–49.

[18] Guney Y, Sari D, Cetin M, Tuncan M. Impact of cyclic wetting-drying onswelling behavior of lime-stabilized soil. Build Environ 2007;42:681–8.

[19] Seco A, Ramirez F, Miqueleiz L, Urmeneta P, Garcia B, Prieto E, et al. Types ofwaste for the production of pozzolanic materials – a review. In: Show KY,editor. Industrial waste. Shanghai: Intech; 2012. p. 141–50.

[20] Xeidakis GS. Stabilization of swelling clays by Mg(OH)2. Factors affectinghydroxy-Mginterlayering in swelling clays. Eng Geol 1996;44:93–106.

[21] Xeidakis GS. Stabilization of swelling clays by Mg(OH)2. Changes in clayproperties after addition of Mg-hydroxide. Eng Geol 1996;44:107–20.

[22] Singh G, Das BM. Soil stabilization with sodium chloride. Transport Res Rec1999;1673:46–55.

[23] Lees G, Abdelkader MO, Hamdani SK. Sodium chloride as an additive in lime-soil stabilisation. Highway Eng 1982;29(12):2–8.

[24] El-Sekelly AM. Characterization of sodium chloride stabilised soils and soil-aggregate mixtures for low-volume roads in Egypt. PhD Dissertation.University of Leeds, England, January; 1987.

[25] Osula D. Laboratory trial of soil–sodium chloride–cement stabilization forproblem laterite. J Transport Eng 1993;119(1):149–58.

[26] Singh G, Ali MM. Comparative study of effectiveness of sodium chloride in soilstabilization for pavement construction. Highway Res Bull 1978;7:1–13.

[27] Davoudi MH, Kabir E. Interaction of lime and sodium chloride in a lowplasticity fine grain soils. J Appl Sci 2011;11(2):330–5.

[28] Wu C, Zhou B, Li ZJ. Test of chlorides mixed with CaO, MgO, and sodium silicatefor dust control and soil stabilization. J Mater Civil Eng 2007;19(1):10–3.

[29] Attom MF, Al-Sharif MM. Soil stabilization with burned olive waste. Appl ClaySci 1998;13:219–30.

[30] Nalbantoglu Z, Tawqif S. Evaluation of the effectiveness of olive cake residue asan expansive soil stabilizer. Environ Geol 2006;2006(50):803–7.

[31] Garcia MA, Chimenos JM, Fernandez AI, Miralles L, Segarra M, Espiell F. Low-grade MgO used to stabilize heavy metals in highly contaminated soils.Chemosphere 2004;56:481–91.

[32] Bruland KW. Trace elements in seawater. In: Riley JP, Chester R, editors.Chemical oceanography. London: Academic Press; 1983. p. 157–220.

[33] Hodge VF, Stetzenbach KJ, Johannesson KH. Similarities in the chemicalcomposition of carbonate groundwaters and seawater. Environ Sci Technol1998;32(17):2481–6.

[34] Alburquerque JA, Gonzalez J, Garcia D, Cegarra J. Agrochemicalcharacterisation of ‘‘alperujo’’, a solid by-product of the two-phasecentrifugation method for olive oil extraction. Bioresource Technol2004;91:195–200.

[35] De la Casa JA, Romero I, Jimenez J, Castro E. Fired clay masonry unitsproduction incorporating two-phase olive mill waste (alperujo). Ceram Int2012;38:5027–37.

[36] Tortosa G, Alburquerque JA, Ait-Baddi G, Cegarra J. The production ofcommercial organic amendments and fertilisers by composting of two-phaseolive mill waste (‘‘alperujo’’). J Clean Prod 2012;26:48–55.

[37] Asociación Española de Normalización y Certificación AENOR, 103500.Geotecnia. Ensayo de compactación. Proctor Normal. Spanish standard UNE103500; 1994.

[38] Asociación Española de Normalización y Certificación AENOR, 103501.Geotecnia. Ensayo de compactación. Proctor Modificado. Spanish standardUNE 103501; 1994.

[39] Asociación Española de Normalización y Certificación AENOR, 103103.Determinación del limite líquido de un suelo por el método del aparato deCasagrande. Spanish standard UNE 103103; 1994.

[40] Asociación Española de Normalización y Certificación AENOR, 103104.Determinación del limite plástico de un suelo. Spanish standard UNE103104; 1993.

[41] Asociación Española de Normalización y Certificación AENOR, 103502. Métodode ensayo para determinar en laboratorio el índice CBR de un suelo. Spanishstandard UNE 103502; 1995.

[42] Asociación Española de Normalización y Certificación AENOR, 103602. Ensayopara calcular la presión de hinchamiento en edómetro. Spanish standard UNE103602:1996.

[43] Akcanca F, Aytekin M. Effect of wetting–drying cycles on swelling behavior oflime stabilized sand–bentonite mixtures. Environ Earth Sci 2012;66:67–74.

[44] Martin-Ramos JD. Using Xpowder: A Software Package for Powder X-RayDiffraction Analysis; 2004 [(GR 1001/104. ISBN 84-609-1497-6)].

[45] Kolias S, Kasselouri-Rigopoulou V, Karhalios A. Stabilisation of clayey soilswith high calcium fly ash and cement. Cem Concr Compos 2005;27:301–13.

[46] Basha EA, Hashim R, Mahmud HB, Muntohar AS. Stabilization of residual soilwith rice husk ash and cement. Constr Build Mater 2005;19:448–53.

[47] Ahmed A, Ugai K, Kamei T. Investigation of recycled gypsum in conjunctionwith waste plastic trays for ground improvement. Constr Build Mater2011;25:208–17.

[48] Ahmed A, Ugai K. Environmental effects on durability of soil stabilized withrecycled gypsum. Cold Reg Sci Technol 2011;66:84–92.

[49] Horpibulsuk S, Phojan W, Suddeepong A, Chinkulkijniwat A, Liu MD. Strengthdevelopment in blended cement admixed saline clay. Appl Clay Sci2012;55:44–52.

[50] Pliego de prescripciones técnicas generales para obras de carreteras y puentes/General technical specifications for construction of roads and bridges. SpanishMinistry of, Development; 1976.

[51] Moore DM, Reynolds RC. X-ray diffraction and the identification and analysisof clay minerals. New York City: Oxford University Press; 1997.

[52] Yilmaz I, Civelekoglu B. Gypsum: an additive for stabilization of swelling claysoils. Appl Clay Sci 2009;44:166–72.